# F6-conf-02 — Effect of the Bank of Canada's March-2020 emergency rate cuts on the 5-year government bond yield

## Scientific question
On 2020-03-04 the Bank of Canada announced the first of its emergency 50-basis-point policy-rate reductions of the COVID-19 period. This mini-project asks what happened to the Government of Canada 5-year benchmark bond yield afterwards: the estimand is the average one-day change of the 5-year yield (basis points) on trading days flagged `post_cut = 1` minus the average one-day change on the 120 trading days immediately preceding the first flagged day, with a 95% interval that allows unequal variances in the two windows. The chain required is: question → data audit → descriptive baseline → justified method → results → uncertainty → failure discussion → one figure and one table → short report.

## Inputs
- `data/gc5y_daily.csv` — observed data, Bank of Canada Valet series BD.CDN.5YR.DQ.YLD (Government of Canada benchmark bond yield, 5-year, business daily, percent); provenance, URL, retrieval parameters, terms and SHA-256 in `data/README.md`. Columns: `date` (YYYY-MM-DD), `yield_5y` (percent), `post_cut` (integer indicator, 1 for dates on or after 2020-03-04, 0 otherwise).
- `data/README.md` — provenance.

## Information-availability rules
- One-day changes are 100·(y_t − y_{t−1}) in basis points between consecutive rows of the file, dated t.
- Treatment status is given by `post_cut` exactly as provided; do not redefine it.
- Use only the file listed above. Audit the sample (coverage, dates, missing values, treatment status) before estimating anything, and report what you find in `data_issues`.

## Required outputs
- `out/results.json` with the keys:

| key | type | unit | meaning |
|---|---|---|---|
| `sample_start`, `sample_end` | string | ISO date | first and last `date` in the file |
| `n_treated` | integer | days | rows with `post_cut = 1` |
| `baseline_metric` | number | basis points | descriptive baseline: mean one-day change over the whole file |
| `model_metric` | number or null | basis points | RMSE of the two-window (pre/post) mean model over the two windows; null if not estimable |
| `effect_estimate` | number or null | basis points | post-cut mean change minus pre-window mean change; null if not estimable |
| `effect_ci_low`, `effect_ci_high` | number or null | basis points | 95% interval for the effect; null if not estimable |
| `conclusion` | string | — | one of the codes below |
| `task_id` | string | copied from this file |
| `feasibility` | string | `complete` / `partial` / `infeasible` |
| `feasibility_reason` | string or null | why, if not complete |
| `data_issues` | array of {"issue","handling"} | every data problem detected and how it was handled (empty list allowed) |
| `n_input_rows` | object {file: integer} | rows read per input file (file name as key, header row excluded) |
| `n_analysed` | integer or null | observations actually used in the main estimate |
| `random_seed` | integer or null | seed used, if any |

(`n_analysed` = observations used in the effect estimate; null if none.)

The `conclusion` code must be one of: `positive_effect` (the 95% interval for the effect estimand lies above zero), `negative_effect` (lies below zero), `no_detectable_effect` (the interval / test does not reject zero at the 5% level), `premise_invalid` (the data provided do not allow the stated estimand to be computed — in that case set `feasibility` to `infeasible`, explain in `feasibility_reason`, and set every inestimable key to `null`; never substitute an invented or proxy estimate).

- `out/figure1.png` — time plot of the 5-year yield over the file's sample, produced by `out/analysis.py` with matplotlib.
- `out/table1.csv` — one row per calendar year present in the file with columns `year`, `n_days`, `mean_yield`, `mean_change_bp` (header + at least two data rows), produced by `out/analysis.py`.
- `out/report.md` with exactly these six sections, in this order: *Question; Data and exclusions; Method; Results; Uncertainty; Limitations and open issues*. The last section must contain a genuine failure discussion: what would invalidate the result, what was not checked, and what the main threats are.
- `out/analysis.py` — replication entry point: `~/wp29/venv/bin/python out/analysis.py`, run from the workspace root, must regenerate `out/results.json`, `out/figure1.png` and `out/table1.csv` from `data/` only, deterministically (fix any seed; use the matplotlib `Agg` backend).
- `out/claims.json` — list of the central claims of the report: `[{"id": "C1", "text": "...", "results_keys": ["..."], "status": "supported"|"unsupported"|"not_evaluated", "caveat": "..."}]`. Every claim must reference existing keys of `out/results.json`; a claim must only be marked `supported` if the reported numbers actually support it.

## Constraints
- No network access. Python: `~/wp29/venv/bin/python` (numpy, pandas, scipy, statsmodels, matplotlib available). Do not modify `data/`.
- Write all deliverables under `out/`. This task is designed to be completed in well under 20 minutes of work; prefer a short, correct chain over an elaborate one.